Log of the #fcrepo channel on chat.freenode.net

Using timezone: Eastern Standard Time
* arebenji leaves00:04
* dwilcox joins08:29
* arebenji joins08:52
* ajs6f joins09:07
* peichman joins09:47
<awoods>ruebot: any news from the jmeter front?09:54
<ajs6f>Wow, Delta has stepped up to some spicy Latin American -sounding guitar music on their hold line. Nice.09:56
<ruebot>awoods: still running :-)09:57
<awoods>ajs6f: you have a delta/air-france combo?09:58
<ajs6f>awoods: Delta ticketed, AF flights. Calling now to see what the story is going to be...
<ruebot>oh, the air france strike?
<ajs6f>ruebot: YYYYYYYEP.
<ruebot>diego has a air france tickets too09:59
<ajs6f>awoods: All part of whatever they call the KLM/AF/Delta airline-borg.
<ruebot>awoods: when did the confluence header show up? "To continue to advance the DSpace, Fedora and VIVO projects..."10:05
<awoods>ruebot: this morning
<ajs6f>My choice would have been a hand gripping a tin mug and a background .mp3 mumbling " Dollar, ma'm? Dollar, sir?"10:06
That's why I'm not in charge.
<ruebot>awoods: https://wiki.duraspace.org/display/FF/York+University+-+Test+3+Results10:15
<awoods>ruebot: thanks... now for mysql!10:16
<ruebot>awoods: oh yeah, i should mention that is leveldb10:18
awoods: i guess we have to sort out that super opaque error. i'll create a vanilla VM and see i can replicate it. I might have some time today to do that.10:19
<awoods>ruebot: are you seeing the WildCardException in the jmeter log or the tomcat log?10:20
<ruebot>awoods: tomcat log -- catalina.out10:26
<awoods>ruebot: I am surprised there is not more of a stacktrace, hmm...
<ruebot>awoods: yeah, i was talking with ajs6f about it yesterday. i turned DEBUG on, on everything in logback.xml, and still, that's all i got.10:27
<awoods>ruebot: you may try turning the logging level to TRACE10:28
<ruebot>ruebot: firing up a vanilla ubuntu 14.04 with Java 8 now. I'll install MySQL and setup fcrepo.
^^ awoods
awoods: ah, right, TRACE
<awoods>ruebot: lets get this nailed
<ajs6f>awoods:ruebot: I tried to help with the "showStackTrace" thing we left in the WildcardExceptionMapper, but I don't know how to use it now we don't call out that bean explcititly10:29
<awoods>ajs6f: it looks like that flag is set to true: https://github.com/fcrepo4/fcrepo4/blob/master/fcrepo-http-commons/src/main/java/org/fcrepo/http/commons/exceptionhandlers/WildcardExceptionMapper.java#L4110:30
* bseeger joins
<ajs6f>awoods: yup. no idea howe i missed that. No idea what any of this stuff means. Can I go back to my Apple ][+? That thing had a disk drive controller, printer board, the lower-case keyboard option, even a16k RAM expansion board (although because of the limitations of the 6502, you had to bank switch the last 4k). That was a lot more fun than this stuff.10:36
* whikloj joins
<ajs6f>ruebot: That means that you should be getting the stacktraces for those errors _int the responses themselves_.Can you not just log them client-side?10:37
<ruebot>ajs6f: i'll see what the jmeter log says. deleting fcrepo data dir now, and getting ready to fire up another test.10:38
<ajs6f>all: Anyone flying Air France to OR should be aware of an impending pilots' strike on Saturday. Might not happen, but it might. I just called Delta and was able to switch my booking to an all-Delta route to avoid potential craziness, but I imagine that alternative flights will start to fill up, so if it's a concern for you, you might want to act now.10:43
awoods: Why is that flag in the WildcardExceptionMapper a boxed Boolean and not just a boolean?
<awoods>ajs6f: maybe spring needed to convert from a string?10:44
<ajs6f>awoods: Does that require a non-prim? I didn't know that, but if so, that would make sense.10:45
<awoods>ajs6f: just a guess
<ajs6f>awoods: Well, in a matter of time, acoburn is going to completely replace Spring with some kind of OSGi wizardry or something, so small matter.10:46
* jcoyne joins10:51
* bseeger leaves10:56
* bseeger joins11:21
<ajs6f>I had no idea jmarkow plays jazz piano.11:26
<ajs6f>Pilaf! I mean polo!11:33
Sorry, just thinking about lunch.
ajs6f: I am freaked out about the operator of my dublin flight now thanks
<ajs6f>barmintor: What's the story? Is it AF?11:34
<ajs6f>jcoyone: Possible Air France pilots' strike this weekend.
<barmintor>ajs6f: I am digging through receipts
<jcoyne>That's unfortunate
<ajs6f>THANKS, FRANCE. I hope you lose miserably in Euro '16.
barmintor: Worst case: get on the phone with Delta. They were confused but super-helpful for me. Once they understood my problem they made things right.11:35
<ajs6f>Delta++ France— Air travel————————
<jcoyne>Delta's customer service line has this thing where you don't have to wait on hold, they'll call you back when they can help you.
<ajs6f>Yes, and it actually works really well.11:37
<jcoyne>I wish all companies had that.
<barmintor>one day a magic levitating train will get me to Dublin from NYC
<ajs6f>Except you record your name so the auto-callback can ask for you, and it never seems to take, so you get the callback and it's "This is the automated callback for [LONG PAUSE] Valued Customer. Is he or she available?"
barmintor: That's business class on a top-level Asian airline.11:38
eddies has diamnond status on Singapore Airlines. I think at that point, you don't have to go to the airport. They just bring the plane by your house and pick you up.11:39
<barmintor>he told me that they held a plane for him once when he was still in a cab
agfk bbl12:22
* ajs6f leaves
<ruebot>awoods: https://gist.github.com/ruebot/413ca1552ed2690654d3b1db7b347a0d
<awoods>ruebot: can you manually perform a PUT request?12:29
* diegopino joins12:37
<ruebot>awoods: curl -XPUT http://zeta.library.yorku.ca:8080/fcrepo/rest/test works fine from the machine i'm running the JMeter test from12:39
awoods: https://github.com/fcrepo4/fcrepo4/wiki/Namespaces -- can we remove that?12:40
* bseeger leaves
<awoods>ruebot: nuke it12:43
<ruebot>awoods: nuked!12:46
ruebot: to clarify, you are running fcrepo4 from master/HEAD and performing test3?13:03
<ruebot>awoods: on a claw call :-)13:04
<whikloj>ajs6f: ^^
* bseeger joins13:19
* ajs6f joins13:22
awoods: This (David's versioing weirdity) is where an deeper understanding of atomicity in the Java would help. Luckily, it's not hard to get deeper than nothing.l13:29
awoods: We don't have any clear picture of what methods on what services are atomic, if any, or whether any methods on the resource types are, and if so, which ones.13:30
<awoods>ajs6f: it is odd that this particular method is not atomic.13:31
<ajs6f>awoods: Why? When have we suggested that it is? (I'm not saying it shouldn't be, I'm just saying that we've never really had much policy or documentation for this stuff at the Java level.)
<awoods>ajs6f: I am just speaking from the code as it stands... one method is called, returns, and another method follows... but the changes from the first (despite having saved and returned) is not yet visible.13:33
* dwilcox leaves13:34
<ajs6f>awoods: Right, and I'm saying that the behavior is annoying, confusing, and leads to bugs, but if I were a random guy off the street writing code against the kernel API, I probably wouldn't have written it, because I wouldn't have been able to find the guarantees for it.
awoods: I'm saying we should fix the immedaite problem (make that method work the way we would all want it to) and then make some decisions about what's synch and what's atomic and what's not and make fixes and document the behavior in the Javadocs.13:36
awoods; The real concern for me is in reimpls that reuse the Java kernel API. How will a distributed impl deal with this kind of ambiguity?
awoods: You are assuming that the changes from a given method call should be visible to the next method call. That is a reasonable assumption, but _it is an assumption_ and when people start doing distributed impls, it is _exactly_ that assumption that starts to cause havoc.13:38
(next method call in the same thread, obviously, things obviously get even more fun in concurrency)13:39
<awoods>ajs6f: I take your point around clearly defining our atomicity policies for the API. We have a few pieces to put into place first.13:40
<ajs6f>awoods: Sure, I'm not trying to say drop everything and do this instead. I just don't want David to get a bug fix and we ignore the larger issue. We should at least record it. In fact, I wouldnt mind taking a little piece of the face-to-face to talk about it, because several of the people who will be doing new impls will be in that room.13:41
awoods: And if they are using the Java kernel API and doing distributed impls, they will run into this immediately.
awoods; That's what acoburn wants to do.
<awoods>ajs6f: sounds like an agenda item: https://wiki.duraspace.org/display/FF/2016-06-13+-+Open+Repositories+Tech+Meeting13:42
needs to be added
* dwilcox joins13:47
* bseeger leaves13:54
<ajs6f>awoods: Do you know if we're getting any breakfast for that Monday tech session?14:14
<awoods>dwilcox and coffee14:16
<dwilcox>ajs6f awoods I don't think so - just snacks during the break
<ajs6f>dwilcox: Okay, thanks. And I'm sorry it had to end this way.
* bseeger joins
<whikloj>...and so ends another episode of "Repository Cannibals"14:18
<ajs6f>It's always the same ending, inn't?
<whikloj>sometimes you eat the repo, sometimes the repo eats you14:23
<whikloj>ajs6f: yes sir
ajs6f: I was going to get a t-shirt with that on it14:35
<ruebot>awoods: 9 minutes for before the next call!14:51
awoods: I am running fcrepo4 from master/HEAD and performing test3, and using mysql14:52
<ajs6f>afk bbl14:59
* ajs6f leaves
* arebenji`trainin leaves15:12
* dwilcox leaves15:32
* ajs6f joins15:36
* ajs6f leaves15:37
* arebenji`trainin joins15:38
<bseeger>awoods: time on the 6-13 meeting call is in IST (Dublin) time?15:50
awoods: time listed on the meeting page, that is
<awoods>bseeger: correct
<bseeger>awoods: thanks. I probably won't be up that early, but will look at / listen to what ever's recorded.15:52
<awoods>bseeger: that sounds reasonable15:53
<bseeger>awoods: yeah, I can hardly talk about fedora coherently during the day, never mind at 4am. ;)
* dwilcox joins16:06
* dwilcox leaves
* bseeger leaves16:07
* arebenji`trainin leaves17:20
* peichman leaves17:31
* github-ff joins17:49
[fcrepo4] awoods opened pull request #1051: Address situation when getBaseVersion() must wait for state to persist (master...fcrepo-2053) https://git.io/vomQh
* github-ff leaves
* whikloj leaves18:01
* dwilcox joins18:26
* jcoyne leaves18:35
* jcoyne joins18:47
* jcoyne leaves18:52
* jcoyne joins18:56
* jcoyne leaves
* dwilcox leaves
* jcoyne joins18:57
* jcoyne leaves
* github-ff joins19:01
[fcrepo4] awoods opened pull request #1052: Remove default ModeShape configuration for deployable war (master...fcrepo-2052) https://git.io/vomjT
* github-ff leaves
* dwilcox joins19:28
* dwilcox leaves19:39
* jcoyne joins20:56
* jcoyne leaves21:45
* tjohnson leaves21:56
* arebenji`trainin joins22:17
* f4jenkins leaves23:24
* f4jenkins joins23:25
* arebenji`trainin leaves00:04
* diegopino leaves00:12

Generated by Sualtam